Lynxmotion Tech Support

The 422 is one of the weakest Hitec servos. The 5485 is a decent servo, but even it is slightly on the anemic side. You could use it if you tethered the power, I.e. no battery on board. The 5645 is the servo of choice for the Scout. Very sorry to say it's $54.99 each. The better choice would be the 5685 which is the same thing only it accepts 7.4vdc directly. It's $56.95. Sorry man...

_________________
Jim Frye, the Robot Guy
http://www.lynxmotion.com
I've always tried to do my best...

Jim,

Thats not really too bad to be honest. I've seen some servos that are $200+

I think i will buy 6 of the 5685 (for the knee and 2 servos either side giving it 'lift' power) as the twist of the hip and ankle are under a lot less momental stress.

Its my birthday next week, so the mrs now knows what shes getting me!!

Cheers Jim,
Matt

These are the servos I'm going to go for as per your advice. Its shocking really how weak the 422 is compared to it, more than 3x the torque!

On the plus side I have 6x422s for sale if anyone wants then for tilt pan heads or something that dont require much power?

Matt

I'm watching this with interest, I'm using a Scout with TowerPro servos (I know they aren't good, but I got them for $5 each and can afford to blow a few, and I know they are jittery as hell, I just want to test some code)... I'm using an Arduino Mega too, directly driving the servos from a Motor Shield.
I have a 7.4v LiPo battery for the servo's (a hefty one) and separate LiPo for the electronics. The plan is to have the weight as low as possible.

I already have it standing and holding position (though it does have 'Parkinsons' due to the servos) and aim to get an IMU and Gyro included with a Kalman filter to stabilise the unit before I start to make it walk.

I guess I need to accept a degree of error due to the servos (they aren't that bad really) but if it stands up on it's own I can start to experiment with COG and walking.
My ideas for movement are a set of subroutines that will accept a parameter, MoveForward 1cm, TurnRight 1 degree etc. Once that is completed I may port the code to a smaller Arduino on an I2C link to the Mega (a co processor shield) which will then operate a sensor tier and possibly wireless to another mobile unit or computer.

The purpose of this unit is to test code and get better at programming (it's been years since I did C in anger and I prefer 6502 assembler), prove my theories and assumptions and ultimately deliver a map of any room it is placed in (wall detection, object detection and placement etc) so I may need more processing than the Arduino can provide, but, it's a goal and I think many projects fail or stall because there wasn't a stated goal to start with.

Beowolf,

Did you ever get very far with your project?

I am just about to order some decent servos for mine, however I may go over to a hexapod chassis as getting this thing to 'gait' properly is nightmare!!

Matt

Arrgghhh...

4th time trying to reply to this posting, the trackpad keeps making the browser jump back and I lose the text!

Ok, so to be brief, yes I did.
Not fantastically far, but it did work.
Unstable because of the servos and lack of integrated gyro and accelerometers but not bad at all.

I'm back here as Im starting this again and wish I could find where I put all the old code and spreadsheets I used to work out the strings for moving.

I started by resetting the servos to a known 'zero' point then depending on what the movement was a table of values was run through the sevo update routine to rotate through 90 degrees, walk forward one step etc...
So I had a set of maybe a dozen callable routines that simply (though somewhat messily to get the memory use down a bit).
I started to get the IMU code in and referred to but got side tracked by the kids, new job and a mobile platform that took over my life for a bit! anyway, I'm about to rebuild the hardware with new metal servo horns, some more powerful digital versions of the MG955 (which isn't a bad servo, if you don't mind the little dancing man instead of one who stands perfectly still) and I'm thinking of using the SCC-32 to store the gaits if I can to reduce memory usage. I have one somewhere...

Did you get the new servos?
How is yours looking now?

It'll take a few weeks to get the horns over from the usa, 9 days apparently! But, cost is everything
then get it rebuilt and start writing the code again. I can't find the original source, just some notes on the table data for each subroutine (having a fixed workshop that the kids and wife don't take over to 'fix' whatever they broke would be great). I'm sure I have a disk with it all on somewhere, its finding it that is the problem!

Once I get going I should take pictures and video and post some code snippets, not that my code is very well written or efficient but it might help...

Did you get further than I did?
This was a help at the time but, it's dissapeared...
viewtopic.php?f=20&t=2827 I wish I could find that zip!
Anyone have a copy???